Deep Learning Automation of Preoperative Radiographic Parameters Associated with Early Periprosthetic Femur Fracture after Total Hip Arthroplasty.

Researchers

Journal

Modalities

Models

Abstract

The radiographic assessment of bone morphology impacts implant selection and fixation type in total hip arthroplasty (THA) and is particularly important in the context of minimizing periprosthetic femur fracture (PFF) risk. We utilized a deep learning (DL) algorithm to automate femoral radiographic parameters and determined which automated parameters were associated with early PFF.Radiographs from a publicly available database and from patients undergoing primary cementless THA at a high-volume institution (2016 to 2020) were obtained. A U-Net algorithm was trained with data augmentation and transfer learning and was optimized to segment femoral landmarks for bone morphology parameter automation. Automated parameters were compared against that of a fellowship-trained surgeon and compared in an independent cohort of 100 patients who underwent THA (50 with early PFF and 50 controls matched by femoral component, age, sex, body mass index, and surgical approach).On the independent cohort, the algorithm generated 1,710 unique measurements for 95 images (5% lesser trochanter identification failure) in 22 minutes. Medullary canal width, femoral cortex width, canal flare index, morphological cortical index, canal bone ratio, and canal calcar ratio had good-to-excellent correlation with surgeon measurements (Pearson’s correlation coefficient:0.76 to 0.96). The metaphyseal morphologic indexes had a moderate correlation (0.50 to 0.63). Canal calcar ratio(0.43±0.08 versus 0.40±0.07) and canal bone ratios(0.39±0.06 versus 0.36±0.06) were higher(P<0.05) in the PFF cohort when comparing the automated parameters.DL-automated parameters demonstrated differences in patients who had and did not have early PFF after cementless primary THA. This algorithm has the potential to complement and improve patient-specific PFF risk prediction tools.Copyright © 2023 Elsevier Inc. All rights reserved.